AI reasons over the network, and makes the network more precise
A single piece of data tells you very little. Structured or not, kernalic builds what came from what and how it connects into a structure, and AI reasons over that network with evidence behind it and feeds what it finds back in to make the network more precise.
What Kernel Ontology does
A person sets the broad frame, and the machine keeps finding further kernels in the data and adds them on top. Only what can cite its source is added, so the structure grows on its own as the data grows.
We define the relationships
What supersedes what, what contains what — we make the relationships between data explicit.
AI reasons over the network
Instead of digging through scattered data piece by piece, AI reasons over the defined network of relationships — answering from the whole web of connections, not a single piece of data.
Every answer carries its evidence
Every relationship records which data, which version and which location it came from, and that evidence follows the reasoning into its results.
It grows more precise on its own
What AI finds while reasoning goes back into the network. The more it reasons, the denser and more precise the network becomes.
Applied domains
The domain the structure is applied to today is intellectual property. SEPPi, for standard-essential patents, runs on it now and keeps expanding to codec, Wi-Fi, Qi and other standards; IPSnap is planned.
One philosophy, two brands
SEPPi first, then IPSnap: both start from the same question: "Where did this data come from, and what does it connect to?"
SEPPi · live · expanding standards
Pinpoints where a standard’s versions and a patent’s claims meet.
IPSnap · planned
Reveals which evidence in a real product meets each limitation of a patent claim.
The kernalic story
kernalic takes its name from the kernel in deep learning. Just as a kernel sweeps one small window across an entire input to pull structure out of it, we start from the broad frame a person sets and let the whole body of data organize and generate the relationships itself, building up the network. We call that method Kernel Ontology, and SEPPi is the brand running on it today.
More about us →Curious what we're building next?
We welcome inquiries about product collaboration, partnerships, and hiring.